Genome Not Designed

[catsittingstill]

One of the analogies Creationists are fond of is to compare genes to software and go on about how an Intelligent Designer must have put together all this complex "software" that "programs" a cell to perform its various functions.

In a stroke of genius, someone has actually compared how "controller" and "middle manager" parts of the genome "call" the workhorse protein-producing parts, and how code intelligently designed (by humans) does the same thing.

It turns out the patterns are very different. Intelligently designed code has a lot of controller and middle manager bits calling a relatively small number of workhorse subroutines.  The genome shows the opposite pattern. As the article explains:

“If you update a low-level function, then you need to update all the functions that use it. That’s doable if you’re an engineer. You just go through all the code. But it’s impossible in biology,” Maslov said.

So changes in low level functions happen when the gene for the low level function is accidentally duplicated.  Then you have a "spare" copy that doesn't need to be maintained to keep the organism alive, and the spare is free to mutate without constraint.  If one (or some combination) of those mutations gives the spare a new, useful, low-level function, great--selection pressure to keep it resumes.  But this leads naturally to an increase of low-level functions that you don't see as much in actually designed code.

Next startling news: Sky Sometimes Blue!

Comments

[maiac.livejournal.com]

You know the Creationists will just say that human-designed code is limited by our mortal human intellects, and God designs code better than we do.

[bedlamhouse]

Let's not confuse designed code with well-designed code. *grin*

In all seriousness, though, I think I recall in my readings (though I have no references) that one "feature" of self-replicating code is that it concentrates on the lower-level functions (spawning code to perform specific functions) rather than efficient use of existing functions or modifications of management code to re-use existing functions in new ways.

One could point out chicken-or-the-egg, in that I am pretty certain that self-replicating code is designed this way due to observations of how natural selection works. However, I also feel this fits with my own personal view of a (non-humaniform, non-interventionist, beyond human understanding) creative force "God" that set the rules in motion and then left them to their own devices for millions of years.

[moshez.livejournal.com]

Just for future reference: http://www.pnas.org/content/early/2010/04/28/0914771107 (PNAS link to the paper)
I don't know why articles about science don't have links to the paper directly. :(

One quotation that irritated me had to do with humans needing to design code more like how the genome is written: hell no. I don't need the equivalent of "appendix" in my code :)